Millimeter-thick microsupercapacitors with linear thickness-scaling of energy and power densities via multilayer electron highway architecture

Abstract

The relentless miniaturization of microelectronics demands energy storage systems with ultrahigh energy and power densities in ultracompact footprints. Microsupercapacitors (MSCs) are promising due to their rapid charge-discharge capabilities, but conventional electrode architectures suffer from a trade-off between energy and power density as thickness increases, exacerbated by electron transport resistance and mechanical instability. Here, we present a paradigm-shifting, truly millimeter-thick (up to 1.2 mm), high-aspect-ratio (7:1), three-dimensional microelectrode architecture that fundamentally decouples electron transport distance from electrode thickness. By roll-to-roll calendering of alternating multilayer current collectors and electrode films, we achieve precisely aligned multilayer structures with total thicknesses exceeding 1 mm, specifically demonstrating a 1.2 mm-thick, 9-layer device in an ultracompact footprint, followed by precision laser engraving to define interdigitated gaps below 180 µm, establishing a parallel electron transport network. Therefore, the architecture enables linear scaling of both energy and power density with thickness, delivering a peak energy density of 1733 µWh cm⁻², comparable to that of 3D micro-batteries, and a peak power density of 153 mW cm⁻², surpassing that of the state-of-the-art microsupercapacitors, alongside exceptional stability demonstrated by 95% capacitance retention after 10,000 cycles at 2000 mV s⁻¹. Roll-to-roll manufacturing combined with laser engraving enables large-area, industrial-scale production of microsupercapacitors, resolving the persistent trilemma among electrode thickness, microscale resolution, and mechanical stability. This advancement delivers a transformative solution for next-generation micro-energy systems achieving high energy density, high power density, and strong mechanical stability.

Introduction

The relentless miniaturization of microelectronics, fueled by the rapid proliferation of the Internet of Things (IoT), implantable medical devices, and autonomous wireless sensors, demands energy storage systems capable of delivering ultrahigh energy and power densities within ultracompact footprint^1–3. Microsupercapacitors (MSCs), with their exceptional power density, long cycle life, and rapid charge-discharge capabilities, are emerging as a critical enabler for next-generation microsystems^4,5. However, their widespread adoption remains constrained by a fundamental limitation: constrained by the very low mass loadings (<1 mg cm⁻²), the energy density of state-of-the-art MSCs, typically below 50 µWh cm⁻², falls far short of the hundreds of µWh cm⁻² required by high-power microelectronics^6–20. This deficit stems from a physical law inherent to conventional electrode architectures, where electron transport distance scales linearly with thickness, forcing an unavoidable trade-off where energy density gains come at the direct expense of power density. Especially at high current densities, electron transport resistance becomes a critical bottleneck, further compromising performance while introducing mechanical instability that undermines structural integrity^21–23. Even worse, as thickness increases, these challenges intensify. Consequently, no existing methods for fabricating MSCs have successfully achieved both high energy density (hundreds of µWh cm⁻²) and high power density (tens of mW cm⁻²) within an ultracompact footprint ( < 0.1 cm²), leaving a critical gap in power solutions for next-generation microelectronics.

The scientific community’s persistent efforts over decades to address this limitation reveal its profound nature as a manufacturing paradox rather than a mere technical challenge. Photolithography-based methods achieve 500 µm thickness but rely on photoresist barriers to prevent short circuits, severely hindering ion transport; or adding binders to the slurry enhances mechanical stability but further reduces electronic conductivity^24–26. 3D printing techniques yield interdigitated gaps exceeding 100 µm, a dimension incompatible with microsystem integration requirements, while simultaneously necessitating non-active additives that degrade electrode performance^27–29. Additive-free inks can boost conductivity but sacrifice resolution, creating interdigital gaps over 1 mm that prevent miniaturization³⁰. Gel electrolyte-filled electrodes improve mechanical stability but introduce elongated electron transport pathways, leading to substantial capacity fading under high current density^31,32. Electrodeposition on metal foams achieves 1200 µm thickness yet delivers only 2.5 mg cm⁻² mass loading, which is insufficient for meaningful energy density gains^33,34. Critically, all these methods share a fatal commonality: they operate within a paradigm where thickness and performance are inextricably linked by physics. As electrode thickness increases, power density inevitably plummets while mechanical instability escalates, a trilemma that has rendered millimeter-level thick microelectrodes commercially unviable despite their theoretical energy density promise^32,33,35,36. The field has reached an impasse where conventional manufacturing methods cannot simultaneously address the extreme requirements of high aspect ratio, high resolution, and industrial-scale production, a quintessential challenge for extreme manufacturing that demands not incremental improvements but a fundamental reimagining of electrode architecture.

While stacking and laser machining have been explored separately, they fail to overcome this paradox alone. Simple stacking exacerbates interlayer resistance, while laser patterning alone cannot shorten the elongated electron pathways in thick electrodes. The true breakthrough requires a paradigm shift towards an architecturally integrated approach that decouples electron transport from electrode thickness^28,32,37,38.

Herein, we designed and successfully fabricated a millimeter-level thick, high-aspect-ratio, three-dimensional (3D) microelectrode architecture that embodies this paradigm shift. This structural design fundamentally alters the direction of electron transport across the entire thick microelectrode, transforming electrode thickness into an advantage rather than a limitation. By leveraging roll-to-roll calendering to construct a laminated structure of alternating aluminum foils and electrode films, we precisely align and embed distributed current collectors into a monolithic, 3D parallel electron transport network. This architecture drastically shortens the electron transport path from any active material to the nearest collector, minimizes internal resistance, and realizes 9 precisely aligned layers (total thickness: 1.2 mm, mass loading: 54 mg cm⁻²) with a 7:1 aspect ratio within an ultracompact footprint of 0.08 cm², effectively decoupling electron pathway length from total electrode thickness. This architectural decoupling is the key novelty of our work. Therefore, both energy density and power density scale linearly with thickness, which is entirely different from previous methods. Subsequent precision laser engraving defines sub-180 µm interdigitated gaps within the 1.2 mm-thick laminated microelectrode, achieving microscale precision within macroscale structures, a hallmark of extreme manufacturing. The synergistic integration of this laser-enabled precision with our re-engineered 3D architecture is what enables the simultaneous achievement of high energy and power density, advancing beyond earlier reports that employed these techniques separately. This two-step process delivers transformative performance, achieving a peak energy density of 1733 µWh cm⁻², comparable to that of 3D micro-batteries and a peak power density of 153 mW cm⁻², surpassing that of the state-of-the-art MSCs. Furthermore, the MSCs maintain 95% capacitance retention after 10,000 cycles at 2000 mV s⁻¹, demonstrating their suitability for long-term power supply to high-power microdevices.

Results and discussion

Electrode design and fabrication

Increasing electrode thickness to boost active material mass loading is an effective strategy for enhancing energy density. However, conventional single-layer current collector architectures suffer from a critical limitation where electron transport distance scales linearly with thickness, severely degrading power density and charge transfer efficiency in MSCs^28,39–41. During the charging-discharging processes, electrons must traverse unidirectionally across the entire electrode thickness (hundreds of micrometers), restricting effective utilization to regions proximal to the current collector (Fig. 1a, the shaded area indicates the effective electrode region). Critically, this thickness-performance trade-off is fundamentally circumvented by our multilayer architecture. When electrodes of identical thickness are supported by multilayer current collectors (Fig. 1b), electron pathways are redirected toward the nearest collector, reducing transport distance by orders of magnitude. Consequently, as additional electrode-current collector-electrode layers are incorporated, energy and power densities scale linearly with electrode thickness without compromising rate capability due to the unchanged electron transport distance. This design simultaneously resolves the long-standing challenge of electron transport limitations in millimeter-thick 3D microelectrodes and addresses mechanical instability through dual-functional aluminum foils^28,32,35,37. Uniformly distributed throughout the electrode stack, these foils function both to establish parallel electron highways that decouple transport distance from thickness and to provide structural reinforcement for high-aspect-ratio (7:1) microelectrodes. The mechanical stability arises from a hierarchical interfacial architecture where, under roll-to-roll calendering pressure, electrode particles embed into conductive binder layers, creating strong multilayer bonding forces between alternating electrode films and binder layers (Fig. 1c). Unlike single-layer current collector configurations reliant on weak adhesion between electrode particles, the multilayer design replaces bulk cohesion with distributed multiple strong interfacial bonding forces, enhancing the mechanical stability and structural integrity. This extreme-manufacturing-compatible process enables sub-180 µm precision patterning within millimeter-scale laminates, eliminating the traditional trilemma between thickness, resolution, and stability.

Fig. 1. Schematic illustration of the design and fabrication of 3D MSCs with high energy density and high power density.

a Schematic diagram illustrating the hindered electron transport in thick electrodes with single-layer current collectors. b Schematic diagram illustrating the order-of-magnitude reduction in electron transport distance for microelectrodes with multilayer current collectors. c The enhancement of interfacial stability by multilayer bonding forces between multilayer electrode films and current collectors with conductive binders. d The lamination manufacturing of large-area multilayer aluminum foils and electrode films by roll-to-roll calendering. e The laser engraving of high-aspect-ratio 3D microelectrodes. f 3D printing of gel electrolyte into 3D MSCs. g The cross-sectional SEM image of laminated multilayer current collectors and electrode films. h The oblique view SEM image of millimeter-level thick, high-aspect-ratio 3D MSCs. i The SEM image of interdigitated electrode surface with 80 µm gaps. j The digital image of large-area laminated multilayer aluminum foils and electrode films. k The digital image of micro-scale 3D MSCs. l 56-Unit 3D MSCs array with 1.2 cm² footprint

Figure 1d–f illustrate the fabrication process of millimeter-thick, high-aspect-ratio 3D MSCs. Under high-pressure roll-to-roll calendering, multilayer aluminum foils and electrode films were alternately laminated to form a thick electrode (Fig. 1d). By adjusting laser engraving parameters, high-aspect-ratio 3D microelectrodes were rapidly and efficiently fabricated (Fig. 1e). Finally, a gel electrolyte was added into MSCs to complete the assembly (Fig. 1f). The successful application of this laser process across devices from single-layer to nine-layer architectures was underpinned by a systematic parametric study. This investigation confirmed that a fixed set of parameters (power, repetition rate, scan speed) could be used to exceed the material ablation threshold, with the number of laser scans serving as the key adaptive variable (see Methods for full details). The required number of scans was found to scale non-linearly with increasing electrode thickness, ranging from approximately 50 scans for a 131 µm-thick electrode to 1000 scans for a 1.2 mm-thick electrode. This thickness-dependent behavior is a direct consequence of the Gaussian energy distribution of the laser beam and the cumulative attenuation through the material stack, which also dictates the transition in interspace morphology from a near-vertical profile in thin electrodes to an inverted triangular cross-section in thicker ones (Fig. S1a, d). This adaptive yet systematic laser process enabled the high-throughput fabrication of all devices presented in this study. Spatially resolved characterization was employed to assess potential laser-induced material alterations. Analysis confirms that the ultrafast laser interaction is confined to a narrow periphery (about 5 to 10 μm) of the ablated gaps, constituting less than 3% of the total electrode area. Within this zone, Raman spectroscopy indicates a reduction in defect density (lower ID/IG ratio), suggesting localized graphitization that may marginally improve interfacial charge transfer. Consequently, the overall electrochemical response remains dominated by the intrinsic properties of the bulk electrode material (Fig. S2). Roll-to-roll calendering laminates 18 alternating layers of aluminum foils and electrode films into structurally robust 1.2 mm-thick stacks with 88 vol. % active material content, the value higher than that of state-of-the-art thick MSCs, enabling high volumetric energy utilization in ultracompact footprints of 0.08 cm² (Fig. 1g, h). Critically, multilayer interfacial bonding sustains integrity during processing, permitting sub-80 μm interdigitated gaps to be laser-engraved within these high-aspect-ratio (8:1) microelectrodes (Figs. 1i, S3), a feat unattainable by photolithography or 3D printing due to mechanical collapse at such scales. By the roll-to-roll calendering process, multilayer aluminum foils and electrode films are alternately laminated to form large-area electrodes, which is compatible with industrial production (Fig. 1j). From the large-area multilayer electrodes, extremely small MSCs with a footprint of 0.08 cm² can be precisely cut by laser engraving, which holds significant importance for the mass production of compact micro-energy devices (Fig. 1k). A critical step in this process is ensuring robust electrical interconnection between the internal aluminum foil layers and the external circuit. This is achieved by preserving the bottom aluminum foil as a conductive lead during laser patterning. Subsequently, a sidewall-deposited aluminum layer, formed by d.c. magnetron sputtering, bridges and interconnects all individual foil layers within the stack. Finally, precise laser ablation removes excess sputtered material to isolate the microelectrodes and define the contact pads (Fig. S4). This synergy of roll-to-roll fabrication and precision laser engraving achieve industrial-scale throughput (56 functional MSCs in an area of 1.2 cm² with each unit occupying 0.35 mm²; Fig. 1l). By simply adjusting laser engraving parameters, it is possible to fabricate MSCs with varying interdigital spacings and widths. The areal capacitance increases with the width of the interdigitated electrodes at low scan rates, whereas larger interdigitated electrode widths lead to a reduction in areal capacitance at high scan rates (Fig. S5). This is attributed to the limited ion mobility at high scan rates⁴². Figure S6 further demonstrates the influence of different interdigital electrode widths on the electrochemical performance of the MSCs.

Mechanical stability of 3D MSCs

The mechanical stability of thick electrodes remains a critical challenge in extreme manufacturing, as weak interfacial adhesion between current collectors and electrode films, as well as among internal electrode particles, often leads to delamination or peeling during charging-discharging processes^37,43. To address this, optimizing the thickness (1, 5, 10 µm) of the conductive binder on the aluminum foil enhances interfacial adhesion and mechanical robustness (Fig. 2a). Cross-sectional SEM and surface analysis (Figs. 2a, S7) reveal that the porous structure of the binder layer facilitates strong bonding with electrode particles during high-pressure roll-to-roll calendering, embedding them into the binder and forming robust interfacial interactions (Fig. 2b). To further quantify the effect of the conductive binder on the interfacial stability between the current collector and the electrode film, peel strength tests were conducted (Fig. S8). Peeling tests on the optimized 5 µm-thick conductive binder layer reveal that interfacial peeling strength surges to 593.5 N m⁻¹ after three peeling tests, which is several times that of commercial supercapacitors and more than ten times the cohesive strength of the electrode film (Fig. 2c). Critically, peeling tests reveal a phenomenon where repeated peeling progressively exposes fresh binder sites, increasing adhesion from 0.69 N (1st peeling) to 11.87 N (3rd peeling) (Fig. 2c), proving that peeling interface occurs within the electrode film (not at the interface between the aluminum foil and the electrode film), a direct consequence of interface strength exceeding the adhesive force between electrode particles. Moreover, from the SEM and elemental mapping images after each peeling test, the changes in the residual amount of electrode particles on the aluminum foil surface also indicate a strong interfacial bond between the aluminum foil and the electrode film (Fig. 2d–f). This mechanism enables the multilayer laminated electrode to withstand industrial roll-to-roll calendering processes and laser engraving without delamination, whereas thick electrodes with single-layer current collector inevitably experience material spalling during the thermal shock of laser engraving due to the weak adhesion between electrode particles. To directly evaluate the mechanical integrity of the full multilayer electrode architecture, we performed peeling tests on complete, integrated electrodes ranging from 1 to 9 layers. Critically, the results (Fig. S9) demonstrate that failure consistently occurred within the electrode material itself (cohesive failure), and not at the interface between the electrode film and the current collector. This provides definitive evidence that the structural integrity of our multilayer electrode is robust. Most significantly, the laminated architecture inverts the thickness-stability relationship. Due to the enhanced interfacial stability, Fig. 2g–k successfully demonstrates the fabrication of 3D MSCs with electrode layers ranging from 1 to 5 layers. Electrode thickness scales linearly from 131 μm to 1200 μm (Fig.S10) while maintaining 88 vol. % active material content and structural integrity, eliminating the trade-off where energy density gains sacrificed mechanical viability. By establishing distributed interfacial bonding as a scalable extreme-manufacturing principle where thicker electrodes strengthen rather than weaken the structure, we have resolved the long-standing issue of mechanical instability that has hindered the commercialization of high-energy MSCs.

Fig. 2. Structural characterization of 3D MSCs with enhanced mechanical stability.

a The SEM images of different binder layer thicknesses and schematic diagrams illustrating the adhesion effects on the electrode films (1 µm, 5 µm, 10 µm). b The SEM image and magnified view of the strong adhesion interface between the current collector and the electrode film. c Peeling force at different binder thicknesses. d The SEM image and the corresponding C (green) elemental mapping after the first peeling test of electrodes with 5 µm conductive binder. e The SEM image and the corresponding C (green) and Al (blue) elemental mappings after the second peeling test of electrodes with 5 µm conductive binder. f The SEM image and the corresponding C (green) and Al (red) elemental mappings after the third peeling test of electrodes with 5 µm conductive binder. g–k The SEM images of 3D MSCs with electrode films varying from 1 to 5 layers

Electrochemical performance: single-layer vs. multilayer current collectors

To illustrate the significant improvement in electrochemical performance achieved by the multilayer current collectors design compared to the single-layer current collector design, a comparative analysis was conducted on electrodes with a thickness of 400 µm (mass loading of 24 mg cm⁻²). The electrodes are labeled as x layer, where x represents the number of current collector layers, and y represents the electrode thickness. As shown in Fig. 3a, the MSCs with a single-layer current collector supporting a 400 µm-thick electrode exhibit an areal capacitance of only 207 mF cm⁻² at a scan rate of 10 mV s⁻¹. In contrast, the MSCs with 4-layer current collectors achieve a significantly improved areal capacitance of 314 mF cm⁻² under the same electrode thickness. This performance enhancement is attributed to the multilayer current collectors architecture, which significantly shortens the electron transport pathways during rapid charge/discharge processes, thereby improving charge transport efficiency and achieving performance optimization. Furthermore, as revealed from cyclic voltammetry (CV) and galvanostatic charge-discharge (GCD) curves, both the MSCs with a single-layer current collector and those with 4-layer current collectors exhibit ideal capacitive behavior across scan rates of 10 to 500 mV s⁻¹ and current densities of 0.5 to 10 mA cm⁻² (Fig. S11). Notably, the CV curves of the MSCs with 4-layer current collectors consistently enclose significantly larger areas than those of the single-layer current collector design at all tested scan rates, indicating superior charge storage capacity. Additionally, as the current density increases, the contrast in electrochemical performance between the MSCs with the 4-layer current collectors and the MSCs with the single-layer current collector becomes more pronounced (Fig. S11). This further confirms that the multilayer current collectors design contributes to achieving both high energy density and high power density in MSCs. Analysis of the GCD profiles in Fig. 3b reveals that at 5 mA cm⁻², compared with the ohmic voltage drop (IR drop) of 116.84 mV of the MSCs with the single-layer current collector, the MSCs with the 4-layer current collectors exhibit smaller IR drop of 45 mV and longer discharge time, indicating markedly larger areal capacitance and smaller internal resistance. Figure 3c demonstrates that at a high current density of 20 mA cm⁻², the capacitance retention rate of the MSCs with the 4-layer current collectors is 41%, significantly higher than the 27% capacitance retention rate of the MSCs with the single-layer current collector. This indicates that the MSCs with the multilayer current collectors exhibit excellent rate performance. The electrochemical impedance spectroscopy (EIS) results in Fig. S12 reveal that, compared to the MSCs with the single-layer current collector, the MSCs with the 4-layer current collectors exhibit a lower equivalent series resistance (ESR) of 13 Ω and a shorter relaxation time of 12.59 s (MSCs with single-layer current collector: ESR of 28 Ω, relaxation time of 31.62 s). These results demonstrate that the multilayer current collectors design enables faster charge transfer kinetics and superior dynamic response, which are crucial for high power energy storage applications. At a high scan rate of 1000 mV s⁻¹, the MSCs with the 4-layer current collectors maintain a capacitance retention rate of 89% after 10,000 cycles, significantly surpassing the 61% retention of the MSCs with the single-layer current collector (Fig. S13). This is attributed to the multilayer strong bonding forces formed by the multilayer aluminum foils and electrode films, which enhance interfacial stability and thereby improve the mechanical stability of the thick electrode. As a result, the structural degradation of the 3D microelectrodes is effectively mitigated under high-rate cycling conditions, providing critical support for the long-term cycling stability of the device.

Fig. 3. Comparison of electrochemical performance between 3D MSCs with the single-layer current collector and multilayer current collectors.

a–c Electrochemical performance comparison of 3D MSCs with the single-layer vs. multilayer current collectors: a CV curves at 10 mV s⁻¹. b GCD profiles at 5 mA cm⁻². c The variation of areal capacitance with current density. d The variation of areal capacitance with electrode thickness in 3D MSCs with the single-layer current collector at different scan rates. e, f The variation of areal capacitance with electrode thickness in 3D MSCs with the multilayer current collectors: e at different current densities. f at different scan rates

Evaluating areal capacitance at high scan rates or current densities is critical for comparing the electrochemical performance of thick MSCs with different current collector designs (Fig. 3d–f). At low scan rates, the areal capacitance increases nonlinearly with electrode thickness. When the scan rate rises to 10 mV s⁻¹, the rate of areal capacitance increase with thickness significantly slows down, which is due to insufficient electron transport kinetics within thick electrodes. Furthermore, at high scan rates ( > 100 mV s⁻¹), the areal capacitance tends to stabilize regardless of the electrode thickness (Fig. 3d). This is attributed to the inability of electrons to rapidly move between the current collector and the top of the electrode, which can be hundreds of micrometers thick, within a short period of time. As a result, certain regions of the electrode become ineffective and fail to contribute to the areal capacitance. This represents a critical bottleneck for millimeter-level thick, 3D MSCs. In stark contrast, the areal capacitance of the MSCs with the multilayer current collectors increases linearly with electrode thickness from low current density (1 mA cm⁻²) to high current density (50 mA cm⁻²) (Fig. 3e). This is attributed to the multilayer current collectors uniformly distributed throughout the thick electrode, which significantly shorten the electron transport pathways, thereby ensuring efficient charge transfer within the thick microelectrode.

To unequivocally validate this linear scaling principle and directly address the feasibility of millimeter-thick devices, the microsupercapacitor with a total thickness of 1.2 mm was fabricated and characterized, firmly placing it in the millimeter-scale regime. This device comprises a nine-layer electrode architecture (900 μm of active material and 300 μm of current collectors). Most critically, as shown in Fig. 3f, we observe a linear relationship between capacitance and scan rate from 50 to 1000 mV s⁻¹. This exceptional rate capability at an extreme thickness of 1.2 mm provides definitive experimental proof that our multilayer architecture successfully decouples electron transport from electrode thickness, overcoming the fundamental limitations that have previously confined MSCs to the micrometer scale.

Electrochemical performance of MSCs with multilayer current collectors

To investigate the effect of electrode thickness (or mass loading) on the electrochemical performance of MSCs with multilayer current collectors, laminated MSCs composed of multilayer aluminum foils and electrode films were fabricated, denoted as x layers, where x represents the number of aluminum foil layers. Notably, for devices with 1 to 5 layers, at a scan rate of 10 mV s⁻¹, the CV curves exhibit a nearly rectangular shape, and the area enclosed by the curves increases with electrode thickness (Fig. 4a), indicating a significant enhancement in areal charge storage capacity. Crucially, this enhancement is achieved without sacrificing volumetric performance. While the volumetric capacitance remained relatively stable from 1 to 5 layers, it increased markedly to 11.25 F cm⁻³ for the 9-layer, millimeter-thick device (Fig. S14), demonstrating exceptional volumetric capacitance^{27,28,42–45}. This pivotal enhancement underscores a key advantage of our architecture: it enables a greater than ninefold boost in active material mass loading at high layer counts without compromising electron/ionic accessibility, thereby simultaneously boosting both areal and volumetric capacitances even at an ultra-thickness of 1.2 mm. This trend definitively resolves the longstanding trade-off between mass loading and charge transport efficiency in thick electrodes. Furthermore, at 5 mA cm⁻², the GCD curves display symmetrical triangular shapes, demonstrating ideal electric double-layer behavior (Fig. 4b). Additionally, the IR drop decreases from 0.21 V (1 layer) to 0.03 V (9 layers) with increasing electrode thickness (Fig. 4b). This reduction in IR drop is attributed to the lower ESR in thicker electrodes, further supported by the EIS results (Fig. 4c). The analysis in Fig. 4d reveals that for MSCs with 5 layers, the areal capacitance is 352 mF cm⁻² at a scan rate of 10 mV s⁻¹, with a capacitance retention rate of 46% at 100 mV s⁻¹, indicating good rate performance. As a supplementary validation of the rate performance of thick microelectrodes, the analysis of Fig. S15 reveals that the MSCs with 5 layers exhibit a capacitance retention rate of 42% at a high current density of 20 mA cm⁻². In contrast, the MSCs with 1layer achieve only a 21% capacitance retention rate at 20 mA cm⁻². This significant improvement in high-rate performance is attributed to the multilayer current collectors design, which reduces the electron transport distance, thereby substantially enhancing the rate capability of the MSCs. Notably, the nine-layer MSCs exhibit a CV curve area and discharge time exceeding nine times those of the single-layer electrode, owing to its trapezoidal structure which results in a proportionally higher mass loading after laser processing^10,46. The Ragone plot in Fig. 4e demonstrates that by increasing the electrode thickness, the energy density of the MSCs is significantly enhanced, while the peak power density increases linearly with electrode thickness. Both power density and energy density exhibit a linear increase with electrode thickness, and this trend persists even at thicknesses up to 1.2 millimeters (Fig. 4f, g). In contrast, previously reported methods of improving energy density by layer-by-layer stacking of electrode materials often severely compromised the power density of MSCs due to the long-range electron transport and increased resistance within the electrodes. As the number of stacked layers increased, the energy density of the MSCs was greatly constrained. After 10,000 cycles at a high scan rate of 2000 mV s⁻¹, the capacitance retention rate remains at 95% (Fig. 4h), highlighting its excellent cycling stability. For mutual verification, Fig. S16 shows a capacitance retention rate of 90% after 10,000 cycles at a high current density of 50 mA cm⁻². The excellent cycling performance maintained under high current densities and high scan rates confirms the outstanding mechanical stability of the 3D MSCs with multilayer current collectors. In the Ragone plot of Fig. 4i, open symbols represent micro-batteries and solid symbols denote MSCs. The 3D MSCs in this work exhibit exceptional performance, delivering a peak energy density of 1733 µWh cm⁻²and a peak power density of 153 mW cm⁻². Notably, the energy density of MSCs we fabricated surpasses that of most reported micro-batteries (including lithium-ion, sodium-ion, potassium-ion micro-batteries, and so on, shown in Fig. 4i and Table S1), which typically exhibit energy densities below 1 J cm^−2,47–51. Furthermore, the peak power density of 153 mW cm⁻² of the MSCs we fabricated also surpasses that of the vast majority of MSCs (Fig. 4i and Table S1). The MSCs with multilayer current collectors successfully combine micro-battery-level energy density with the exceptional power density characteristic of MSCs.

Fig. 4. Electrochemical performance of MSCs with multilayer current collectors.

a–e The electrochemical performance of MSCs with varying layers of aluminum foils and electrode thickness: a CV curves at 10 mV s⁻¹. b GCD profiles at 5 mA cm⁻². c EIS results. d Areal capacitance at different scan rates. e Ragone plots with different layers. f Energy density as a function of electrode thickness and mass loading. g Power density as a function of electrode thickness and mass loading. h Cycling performance at 2000 mV s⁻¹. Inset showing CV curves of the first and the 10000th cycles. i Comparison of the Ragone plot of this work with other reported MSCs and micro-batteries

The unprecedented integrated performance, achieving record-high areal and volumetric energy densities without sacrificing power density or rate capability, ultimately stems from a fundamental architectural innovation: the intentional trade-off of mass for unparalleled charge transport efficiency. While the incorporation of additional aluminum current collectors increases the mass of the device, it is this very design that decisively overcomes the electron transport bottleneck that has plagued conventional thick-electrode architectures. By creating parallel, shortened electron pathways, the multilayer structure ensures efficient utilization of the active material at ultra-high loadings, directly enabling the linear scaling of performance with thickness demonstrated in Fig. 3e, f, and 4f, g. This strategic design trade-off is overwhelmingly justified by the monumental gains in the most critical metrics for miniaturized electronics, namely areal and volumetric performance, as unequivocally validated by the Ragone plot in Fig. 4i.

Integrated applications of MSCs

The advancement of autonomous microsystems, such as micro-actuators and micro-robots, is critically constrained by the lack of on-chip power sources that can deliver high operating voltages ( > 10 V) within a minimal footprint⁵². Serial connection of individual micro-supercapacitors (MSCs) is the essential strategy to achieve high voltage. Initial experiments confirm the fundamental feasibility of this approach. As shown in Fig. 5a–d, two MSC units (footprint: 0.08 cm² per unit) were successfully connected in series and parallel, demonstrating the anticipated doubling of output voltage and discharge time, respectively. In Fig. 5d, the discharge time of the connected devices is about 20% longer than twice the discharge time of a single unit. The observed about 20% increase in discharge time (compared to twice the time of a single unit) is a characteristic and advantageous outcome of parallel integration, primarily due to a reduction in the effective internal resistance and the associated mitigation of polarization losses.

Fig. 5. Integrated applications of MSCs.

a Comparison of CV curves between a single unit and two series-connected 3D MSCs at a scan rate of 50 mV s⁻¹. b Comparison of CV curves between a single unit and two parallel-connected 3D MSCs at a scan rate of 50 mV s⁻¹. c Comparison of GCD profiles between a single unit and two series-connected 3D MSCs at 1 mA cm⁻². d Comparison of GCD profiles between a single unit and two parallel-connected 3D MSCs at 1 mA cm⁻². e Schematic diagram of a high-voltage micro-supercapacitor. f, g CV and GCD curves of the device. h Voltage density of micro-supercapacitor arrays. i Energy density of micro-supercapacitor arrays

In microsystems, serially connecting multiple individual micro-supercapacitors within a compact footprint faces two critical challenges: (1) establishing long-term reliable interconnections in a confined micro-scale area is extremely difficult, and (2) electrolyte crossover between adjacent units becomes a severe issue, leading to failure and raising safety concerns. These challenges make the final assembly unsafe, complex to manufacture, and ultimately impractical for robust microsystems. Therefore, the development of a safe, simple, and effective integration method for high-voltage MSCs that also maintains high energy and power density is of paramount importance and highly anticipated, despite its significant difficulties. Our work directly addresses this critical application bottleneck by demonstrating a monolithic integration strategy that bypasses these limitations entirely.

Our work directly addresses these challenges by enabling monolithic fabrication of series-connected MSC arrays through laser micromachining. By reducing the unit cell footprint to just 0.01 cm² and integrating interconnects directly into the device architecture, we achieve compact, robust, and scalable high-voltage power sources. Crucially, the structured grid design effectively prevents electrolyte mixing, even in densely packed configurations (Fig. 5e). We constructed two series-connected micro-supercapacitor arrays, each consisting of nine units: one based on conventional single-layer electrodes and the other on vertically stacked 9-layer electrodes (height up to 1.2 mm). Notably, under matched operating voltage (27 V) and current density, the 9-layer stacked array exhibits more than a 9-fold increase in CV integral and discharge time compared to the single-layer array (Fig. 5f, g). This enhancement arises from the tapered 3D interdigitated electrode architecture formed during laser processing, which enables an effective mass loading exceeding nine times that of a single-layer unit (Fig. S1). Ultimately, our integrated high-voltage series device achieved a high voltage density of 55.1 V cm⁻² and an energy density of 379 µWh cm⁻², calculated based on the total footprint of the entire device, including the area of all individual units and the spacing between them (Fig. 5h, i)^53–59. To the best of our knowledge, this energy density value is two orders of magnitude higher than those reported in existing literature.

Conclusion

In summary, we have developed a revolutionary design for MSCs that overcomes the inherent trade-off between electrode thickness and performance by leveraging multilayer aluminum foils as dual-function materials. These uniformly distributed aluminum foils across the thick electrode provide parallel electron transport pathways that decouple electron transport distance from electrode thickness, enabling linear scaling of both energy and power density with thickness. Simultaneously, the strong interfacial bonding between aluminum foils and electrode films ensures mechanical stability, even in millimeter-level thick, high-aspect-ratio microelectrodes. By integrating roll-to-roll calendering and precision laser engraving, we achieve a 10-layer laminated microelectrode with a total thickness of 660 µm, an 8:1 aspect ratio, and an ultracompact footprint of 0.08 cm². This design delivers peak energy density of 2.95 J cm⁻² (820 µWh cm⁻²), comparable to 3D micro-batteries, and peak power density of 68 mW cm⁻², surpassing state-of-the-art MSCs. The 95% capacitance retention after 10,000 cycles at 2000 mV s⁻¹ underscores the robustness of the design under extreme conditions.

Crucially, the roll-to-roll manufacturing process enables large-area, high-throughput fabrication, bridging the gap between lab-scale experimentation and industrial scalability. The sub-80 µm interdigitated gaps within millimeter-thick electrodes represent a paradigm shift in extreme manufacturing, achieving microscale precision in macroscale structures. This breakthrough not only resolves the long-standing trilemma of thickness, resolution, and stability but also sets a new benchmark for high-performance, miniaturized energy storage systems. The modular architecture of the multilayer design further demonstrates its versatility, as it will be adapted to graphene, MXene, and other advanced electrode materials, opening new avenues for next-generation microsystems requiring simultaneous high energy density, power density, and mechanical stability.

Methods

Peeling test

In this work, a 180-degree peel test was conducted, as shown in Fig. S5. All experiments were performed on a tensile testing machine. The tape used for the peel test was 3 M VHB 4920, and the peeling speed was kept constant at 50 mm min⁻¹. The test electrode strips used had dimensions of 19 mm × 60 mm, and the width of the 3 M tape was slightly wider than the test electrode strips at 20 mm. The peel strength was calculated based on the average peel force along the length of the electrode strip, and the peel strength was determined using the following formula:

$$G=\frac{F_{\mathrm{peeling}}}{\mathrm{\omega }}$$ 1

In Formula (1), G represents the interfacial peel strength, F_peeling denotes the peel force, and ω indicates the width of the electrode strip.

Preparation of activated carbon films

Activated carbon particles (YEC-8B), multi-walled carbon nanotubes (MWCNT), and polytetrafluoroethylene emulsion (PTFE, solid content 60%) were prepared in an 8:1:1 ratio. In a beaker, 10 ml of deionized water was added, followed by 1.67 g of PTFE emulsion. Then, 50 ml of anhydrous ethanol was added to break the PTFE emulsion. Subsequently, 8 g of activated carbon particles and 1 g of multi-walled carbon nanotubes were added sequentially and mixed thoroughly. The mixture was placed in an oven at 120 °C for approximately 1 h. When the liquid in the beaker evaporated, and the mixture formed a dough-like consistency, a self-supporting flexible activated carbon electrode membrane was prepared using roll-to-roll calendering.

Preparation of highly conductive binders

Multi-walled carbon nanotubes (MWCNT), CMC, and SBR were prepared in a ratio of 10:2:3. Specifically, 10 g of MWCNT and 2 g of CMC were mixed thoroughly in a blender, followed by the addition of 200 ml of deionized water. The mixture was then stirred at a high speed of 2000 rpm for 1 h. Subsequently, 7.5 g of SBR (with a solid content of 40%) was added, and the mixture was stirred at a low speed of 200 rpm for 10 min.

Preparation of current collectors with double-sided coated conductive binders

A highly conductive adhesive binder was applied onto the aluminum foil using a blade coating method at a coating speed of 600 mm per minute. After baking at 80 °C, the same parameters were used to coat the conductive binder on the other side of the aluminum foil.

Roll-to-roll calendering process

The calendering process was conducted at a speed of 1.2 m min⁻¹, with the applied pressure ranging from 10 to 40 MPa depending on the initial electrode thickness. The aluminum foil current collectors, dual-side coated with conductive binder via doctor-blading, were alternately stacked with activated carbon electrode films. Bonding forces between the electrode films and current collectors were achieved through roll-to-roll calendering, resulting in robust interfacial adhesion.

Laser engraving for fabrication of 3D MSCs

The laser patterning was performed using a FemtoYL-40 femtosecond laser system (Anyang Laser), which delivers pulses with a duration of 484 fs and a central wavelength of 1033.09 nm. The system is equipped with an F-theta lens with a focal length of 100 mm, and the focal plane was precisely calibrated to be 2.15 mm below the internal reference plane to achieve optimal focusing at the material surface. Key processing parameters include a laser power of 1.2 W, a repetition rate of 50 kHz, a scanning speed of 600 mm s⁻¹, and a scan count varying from 50 to 1000 depending on the electrode thickness.

Preparation of gel electrolytes

First, PVDF-HFP (1 g) was dissolved in acetone (8 ml), and a homogeneous transparent solution was formed after stirring. Subsequently, 1.6 g of EMIMBF₄ was added, and after continuous stirring for 24 h, the PVDF-HFP/EMIMBF₄ gel electrolyte was obtained.

Electrochemical characterization

Cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), and Electrochemical impedance spectroscopy (EIS) measurements were performed in a two-electrode system using an electrochemical working station (Versastat 3, Princeton Applied Research, USA). CV examined at the scan rates of 1–2,000 mV s⁻¹. EIS were carried out in the frequency range of 0.01–100 kHz with a 5 mV ac amplitude.

The capacitance values of the full device were calculated from the CV data by the following Eq. (2):

$$C_{\mathrm{device}}=\frac{{\int }_{{\mathrm{V}}_{\mathrm{i}}}^{{\mathrm{V}}_{\mathrm{f}}}I\mathrm{dV}}{\vartheta \mathrm{\Delta }\mathrm{U}}$$ 2

Where C_device represents the total capacitance of the full device (including the active electrode area and the gaps between the interdigitated electrodes), ϑ denotes the corresponding scan rate in the CV test, ∆U indicates the potential window, V_f and V_i represent the final and initial voltage, I is the voltammetric discharge current, respectively.

The areal capacitance (C_areal) were calculated based on the area of the full device (including active electrodes, interdigital gaps, and pins) from the discharge curves of CV curves and GCD profiles by the following formula (3) and (4):

$$C_{\mathrm{areal}}=\frac{C_{\mathrm{devive}}}{{\mathrm{A}}_{\mathrm{device}}}$$ 3

$$C_{\mathrm{areal}}=\frac{I\mathrm{\Delta }t}{{\mathrm{A}}_{\mathrm{device}}\mathrm{\Delta }\mathrm{U}}$$ 4

Where C_areal represents the areal capacitance of the full device, A_device denotes the footprint of the full device, including the active electrode area and the gaps between the interdigitated electrodes, and C_devive indicates the total capacitance of the full device obtained from Eq. (2). It is worth noting that C_areal, A_device, and C_areal are all calculated based on the footprint of the full device. I is the voltammetry discharge current. ∆t refers to the discharge time (from GCD profiles). ∆U indicates the potential window.

The areal energy density (E_areal) of the device was calculated by the following Eq. (5):

$$E_{\mathrm{areal}}=\frac{1}{2}\times C_{\mathrm{areal}}\times \frac{{\mathrm{\Delta }\mathrm{U}}^2}{3600}$$ 5

Where E_areal represents the areal energy density of the device, C_areal is obtained from formula (3), ∆U indicates the potential window.

The areal power density (P_areal) of the device was calculated by the following Eq. (6):

$$P_{\mathrm{areal}}=\frac{E_{\mathrm{areal}}}{\mathrm{\Delta }t}\times 3600$$ 6

Where P_areal represents the areal energy density of the device. E_areal denotes the areal energy density of the device obtained from Eq. (5), ∆t refers to the discharge time (∆t = ∆U/ϑ).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflict of interest

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41378-026-01266-5.